The second version of the statistics of live log entries

The previous study was rightly criticized about the lack of evidence in the source data, which was vividly seen in the example of the “Ufa photo” tag. This time I corrected the situation and repeatedly increased the sample.

Initial data

The study involved 10,000 user diaries. There were missed the first hundred users, after which two users were selected from each rating page. Each user downloaded all the records up to 2006 . The exceptions were 18+ and subzamochnye posts. Records, in turn, have been extracted headlines, tags, text without HTML markup, the number of images and comments. The result was 1,777,308 records.

Enjoy watching!

Records

Header availability

')

Header length in characters

Character length

As suggested by avenu , LiveJournal is very similar to Twitter by the prevailing number of characters in posts.

Hours and days of the week

Number of user records

Looks like a conspiracy. A cursory review of the diaries showed no pattern.

Tags

The length and number of tags

Popular tags

pepelsbey , lytdybr is!

Comments

Number of comments to the post

The length of the text to the number of comments

The average length of the text to the number of comments

Conclusion

The increase in the number of entries significantly affected only popular tags and had almost no effect on the other parameters.

Thanks for attention. Waiting for your suggestions, criticism and comments.

UPD: User logs with 55 entries:
13whitemice , 55thairborngirl , a-mne-eshe , a-sebrov , akmych , al-re , ally-of-sunbeam , anton-platov , b0risl0dkin , bazil-t , beobachter , blog-knockknock , boriansky , brom-termit , catrin- flame , curious-ja , cybercool , da4 , dj-nicks , djrediska , dr-bass , dugla , dyxlesska , echarri , ekateriana , eklery , ennochka , ermolaev-vlad , escaldo , estetika-nice , fabyla , father-kot , geyzer76 gizir , gonish , green-tiffani , gyqyv , hmixa , iliora , jazz-fun , jelka3 , john-scar , k0mpas , karibus , krysia-i-basia , kushka , lagoun , lazutkina , light-tm , loony2004 , love-spring , magnumx , makova547 , malone-xbit , mariri , mashki , mia312 , minorland , more-produkt , mozgovik , nankin , new-zebra , nikita-avanti , oksk , ovine , pastsimple , pavel-lv , peshi-eshe , poignant-art , pugachevsky , roketa , ryzha-sonya , samaposebe99 , sank-a , saule-marsault , schattenphonix , seligoroff , sergik1977 , servinn , sevavladimirov , shtefanesko , sklyankin , snow-cat , stas-y , stei nboom , suhaverhi , svetik-sh , tamikori , tipo-femmina , tri- 4tyre , turobei , uberlastung , ulianich , users , vale4ka-babo4ka , vernon- dimirest
I completely forgot that not all user records were retrieved, but only until 2006. Therefore, the probability of bots becomes quite low. Most of the diaries are live: there are entries for November and December (I collected data in October). Therefore, it is not clear that this may be. It is necessary to collect similar statistics for some other resource (habru?).